Dear Statalist:
I have run into an odd situation when using ivreg2. Upon adding a
particular variable (X) to a regression model as an included instrument
(i.e., exogenous regressor), both the centered and uncentered R^2 decline.
The variable X is significantly related to the outcome Y in the model. When
I estimate the same model using 2SLS, the R^2 also declines when X is added
as an exogenous variable. However, the centered R^2 is much greater in
magnitude than when GMM is used. Any thoughts as to why this might be
occurring? Is it something I should be worried about? Are the R^2
statistics reported by ivreg2 valid in such a situation? If not, what
should I report?
Thanks.
--
James W. Shaw
Research Associate
College of Pharmacy
The University of Arizona
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
